Vol. #127: Epistolary Imperative and Extensionalism/ De-Extinction/ Forever Chemicals No More?/ How to Build a GPT-3 for Science
Massimo Portincaso
Founder & CEO at Arsenale Bioyards, Industrial Romantic and Antidisciplinarian Stoic
Epistolary Imperative and Extensionalism. The new Lux Investor letter is out, and, as always, is incredibly well written and full of insight. I have to admit that among all deep tech investors, LUX Capital is the one that for me really stands out, and the one closest to my ideal of an investor deeply?understanding the space and operating with the level of insight and technical understanding necessary to be successful.
No surprise then, that I was very excited to immerse myself in the reading of the document, which I encourage doing. While long, it is worth the effort. I will not attempt to summarize it, as I would not pay justice to it, and if you want a high-level summary, here is a good twitter trail.
The letter triggered two reflections from my side, around two of the points made in the letter.
Epistolary Imperative. The letter starts with the epistolary imperative of the Lux letter: the pursuit of meaning in a stochastic universe. This is what I like about Lux, as they managed to free themselves from linear thinking, and are always going for the bigger picture. If you read the letter, I do agree with them that the times ahead are full of opportunities, and that what might seem to many a time of doom and gloom, can and will become a time for incredible opportunities.?
But to be able to capitalize on this opportunities, you need to be able to see things the right way. This sentence from the letter summarizes it perfectly “Our perspective is to keep perspective: rip from the trendlines and not the headlines. Read present headlines and it may seem like the worst of times – read past history and it might seem like the best of times. At present we fell conditionally optimistic – conversely contrary to others who are either diffidently in denial that anything has changed for the worse or definitely pessimistic that everything has. Our take: what may feel like a climax of chaos may yet yield a denouement of detente”.?
I really like this perspective and completely agree with the take. I could have not expressed better myself. I would only add one comment. It is not going to be easy, it is going to be painful, particularly for the next two (possibly three) years. The way I am starting to see this, is that we are going to have a couple of years with war like conditions for the economy, and the same way WWII provided much of the innovation and the foundation that powered post war growth, this period can lead to a major transition in the way the world operate, literally.?
What I mean with war like conditions is that we have many fronts to fight, energy transition, climate, geopolitical tensions, and technology will play a very important role on all of these fronts and will be essential to win. This will lead to extraordinary operational and economic conditions, similar to what happens during bellicose conflicts. Ultimately the stakes are extremely high, with the survival of the planet at stake.?
The other point that made me think was the one about Extensionalism. The core idea is that we, as humans, are being “extended” in our capacities and capabilities, through extensions of our cognitive and sensing capabilities powered by technology. The point is valid, and clearly is one of the element of Lux’s investment thesis. I would like though, to push the thinking even further, because the core insight behind Extensionalism is very powerful and is at the core of what deep tech is.?
What is making this point in time so exciting, despite all the negative news we are hearing, is that, for the first time, we are in a position to manipulate matter (and energy) with atomic precision (both at the organic and inorganic level) and, through accurate sensing and actuation leverage cognition and computation to manipulate it, achieving breakthroughs that were unthinkable, and fully explore the option space provided by Nature. If you want, we are entering the era of “Nature-Extensionalism”, which is what I have been calling Nature Co-Design.??
PS: Apologies if the this week’s layout is suboptimal. I managed to forget home my laptop before a 10 days trip… and am now operating from an iPad, which is great, but not the same as a laptop
The last of the thylacines - AKA Tasmanian tigers - “died out in the 1930s.” Now, the University of Melbourne has received a $5M gift?to help bring the marsupial “apex predator” back to life. UM researchers have partnered with Colossal?- a biotech “de-extinction?company” - and plan to use the “previously sequenced genome?of a juvenile specimen” as “a complete blueprint on how to… build a thylacine.” If successful, the goal is to eventually reintroduce the Tasmanian tiger into the wild.
Colossal also plans to use its genetic engineering tech to ”recreate the woolly mammoth and return it to the Arctic tundra” within six years.
News items:
The ?Inflation Reduction Act?was hailed?as “the most important climate action in US history.” It provides significant subsidies and funding for ?carbon capture and storage?(CCS). In this NY Times op-ed, the founders of the ?first privately funded CCS startup?in the US say “clean power from carbon capture and sequestration died with the success of renewable energy” and call for CCS tech to be “buried deep underground.”
Often known as “forever chemicals,” ?PFAS?are “found in a spectrum of products and contaminate water and soil around the world.” After searching for ways to destroy them for years, chemists have come up with a?relatively simple method?of rendering “PFAS molecules harmless by mixing them with two inexpensive compounds at a low boil.”
According to ?Tasha Stoiber, Senior Scientist at ?EWG: “This research is desperately needed, [but] the reality of the situation is that there is really no magic solution right now other than undertaking the hard work of recognizing just how difficult the [PFAS] problem is and turning off the tap so that we don’t make it any worse.”
News items:
In 2019, Google claimed that its quantum computer, ?Sycamore, had ?passed quantum supremacy?by performing an “abstruse calculation” that would take a supercomputer “10,000 years” in 200 seconds. Chinese scientists recently completed the “computation in a few hours [using] ordinary processors. A supercomputer, they say, could beat Sycamore outright.”
Why isn’t there a ?DALL·E 2?or ?GPT-3?for science? Graphic designers and writers around the world live in fear of impending obsolescence, but scientists appear immune to automation. Are they?
“Scientific research is the world’s most valuable content,” but it’s also the “least accessible and understandable.” It often resides behind a paywall, publisher, or other gatekeepers, keeping it largely inaccessible. ?Scite?co-founder ?Josh Nicholson?argues that by establishing new relationships with publishers like ?Wiley?and gaining access to their resources, a GPT-3 for scientific research is within reach.
News items:
With over ?$1.4B invested in climate tech?in Q2 2022 alone, carbon dioxide removal (CDR) and ?Direct Air Capture(DAC) startups?are awash in capital. But without “sufficient oversight” or governance: “The really uncharted territory is what CDR means for environmental and social justice.”
OpenAI ?recently announced?that DALL·E 2 is being released as an open beta. Initially, users will be given some free monthly credits, with the option to purchase more as needed. “More importantly, [OpenAI] will allow using the generated images commercially,” as will competitors like ?Midjourney?and ?Stable Diffusion. According to ?The Algorithmic Bridge: “Digital AI-savvy artists that know how to leverage these models to enhance their toolkits and improve their creations are notably happy.” But a growing number of human artists - like ?David O’Reilly?and ?Karla Ortiz- are not.
Many take issue with DALL·E 2 and other image-generation models being trained on the work of living creators without permission or compensation. Are these “opaque” “black boxes” copying, reproducing, or plagiarizing? And what can be done about it when such models - and the corporations that own them - aren’t subject to any form of regulation?
News items:
Noah Bushnell founded Atari, Chuck E. Cheese, and gave Steve Jobs his “first corporate role at 19.” Now he’s taking on ?The Future of Education?with a forthcoming book and the ?ExoDexa?gamified learning platform.
In 2021, Google “filed over 600K reports?of child abuse material and disabled the accounts of over 270,000 users as a result.” One such user, a software engineer and stay-at-home dad named Mark, got caught up in Google’s “algorithmic net designed to snare people exchanging child sexual abuse material” when he took photos of his toddler’s “swollen” and painful “groin area” which his wife uploaded to their health care provider at a nurse’s request.
Two days later, Mark received a notification from Google that his account had been suspended because of “harmful content” that was “a severe violation of Google’s policies and might be illegal” and provided a ?list of possible reasons. Ultimately, Google’s judgment - and refusal to consider reversing it - “cost [Mark] more than a decade of contacts, emails, and photos, and made him the target of a police investigation.”
News items:
A ?recent UK study?of almost 40K gamers “found little to no evidence for a causal connection between game play and well-being.”